From: ibuprofin@painkiller.example.tld.invalid   
      
   On Sun, 10 Feb 2013, in the Usenet newsgroup alt.os.linux.mandriva, in article   
   , Bobbie Sellers wrote:   
      
   >Jim Beard wrote:   
      
   >> Adam wrote:   
      
   >>> Sometimes I have no idea where spell checkers get their basic   
   >>> list from. I think some of them must be decades old.   
      
   That's a big old "that depends" - Mentioned in my other reply, I'm   
   using two lists, one from the MOBY project (which was DOS shareware   
   as I understand it) from 1988, and the Webster's Second International   
   originally copyrighted in 1934.  Much of what I've seen has been   
   created by inhaling huge quantities of text much as a classic   
   dictionary creator would do.   
      
   >> When spell-checkers first came into general use, I thought they   
   >> were great. I just kept putting words the spell-checker did not   
   >> know into the list it used, and all was great. Until a new OS or a   
   >> new wordprocessing program came along, and all that effort was in   
   >> the used bit bin.   
      
   Let's qualify some of that - in UNIX, I'm used to the spell checker   
   looking at a "word list" - currently in /usr/share/dict/ and (here in   
   the USA) consisting of words in ASCII, one per line.  There's often a   
   DSW going on about the lists, and problems with copyrights and such.   
   From the LSM:   
   =========================   
   Begin2   
   Title         /usr/dict/words for Linux (linux.words)   
   Version       2   
   Desc1         This is word list containing 45402 words.  Great care has   
   Desc2         been taken to be sure that this word list is free of   
   Desc3         copyright.  This list is suitable for English language   
   Desc4         spelling checkers and as a target for look(1).   
   Author        Rik Faith   
   AuthorEmail   faith@cs.unc.edu   
   Maintainer    Rik Faith   
   MaintEmail    faith@cs.unc.edu   
   Site1         ftp.cs.unc.edu   
   Path1         /pub/faith/linux/utils   
   File1         linux.words.2.tar.gz   
   FileSize1     140k   
   Site2         tsx-11.mit.edu   
   Path2         /pub/linux/docs   
   Site3         sunsite.unc.edu   
   Path3         /pub/Linux/libs   
   CopyPolicy1   Free of non-commercial restrictions.   
   CopyPolicy2   Free for personal, educational, and research purposes.   
   Keywords      dict, dictionary, words, wordlist   
   Entered       Sun Oct 10 19:02:47 1993   
   EnteredBy     Rik Faith   
   CheckedEmail  faith@cs.unc.edu   
   End   
   =========================   
      
   but most distributions have far larger lists, and multiple languages.   
      
   Those words that _you_ are adding go into a "personal" dictionary or   
   word-list, and this is a dot-file in your home directory - examples   
   such as "~/.ispell_english", "~/.hunspell_en_US" or "~/.aspell.en.pws"   
   or similar.  Easiest way to find the personal list is to use the spell   
   checker to check a document with an intentionally mis-spelled word,   
   tell the spell checker to add that word to the dictionary, and then   
   use 'ls -ldart .[a-z]* | tail'   
      
   I'm not using windoze, and my word-processor is "/bin/vi", so I can't   
   answer for what a real word processor might use/do.   
      
   >        Does anyone know how to do this, back up the work   
   >and add the modified dictionary to a new installation?   
      
   This _MAY_ depend on the distribution - I've seen them using several   
   different spell-checker programs, but as mentioned they tend to be   
   using wordlists in /usr/share/dict/.   Some spell checkers allow you   
   to add word-lists, and the hoop-jumping is described in the man-page.   
      
   >        I have been thru the same hassle myself as my   
   >posts to several list use specialised jargon and for   
   >example medical terminology.   
      
   Yup - that's true with many technical fields.   
      
   >Adding to the dictionary seems to be broken.  There seems to be no   
   >way to permanently add words in proper spellings to the things.   
      
   'ls -ldart ~/.[a-z]*'  and see if you can identify the "personal"   
   word list.  These are (in the US) typically ASCII text, one word per   
   line, and as such, generic across distributions and spell-checkers.   
   the FILENAME may be something weird, but the contents are normal.   
   Because I'm using a networked file-system, my home directory is on   
   the file-server, and that is unchanged as I may replace computers,   
   never mind distributions.  My current file has been around for a   
   while, as it still has "Sunnyvale" and the street I used to live on   
   there, neither of which were in the "Webster's Second International"   
   (and I've been here for almost 17 years).   
      
           Old guy   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)